What is Card Sorting?

A/B testing is often the go-to play of digital optimization programs, but it comes with a serious limitation that many teams ignore: It can’t generate new ideas on its own.

You see, A/B testing is evaluative research. It can tell you if one option works better or worse than another option. However, it can require significant time and traffic before you can make informed decisions based on the data collected.

When you need a direction to start testing, it’s important to use generative research: Exploratory user research that focuses on discovering and understanding user needs, behaviors, motivations, and expected experiences.

Card sorting is one of the most powerful generative research methods we have at our disposal and a key part of a full research program. It’s a tool we use with clients because it can drill deep into what users expect from a digital experience.

In this article, we’re going to dive deep into card sorting. This research method is a powerful tool every optimizer should have in their toolkit.

What is Card Sorting?

Card sorting is a research technique used to understand how users organize and categorize information. Essentially, you present participants with a set of cards, each representing a piece of content or functionality, and ask them to group these cards in a way that makes sense to them.

This process reveals how people naturally think about and structure information. It lets you uncover insights into how users might intuitively organize menu items, product categories, or any other structured content on your site. Ultimately, this helps you design a website or app that aligns with their expectations.

3 Types of Card Sorting: Open, Closed, and Hybrid

There are three main types of card sorting methods. They all help you gain a deeper understanding of how users categorize information, but they have their own use cases. Choosing the right type comes down to the main objective of your study.

Open Card Sorting

In open card sorting, participants organize cards into categories and then label the groups themselves. Researchers do not provide any guidance. This is a generative research method in that it helps define the categories rather than evaluating existing ones.

Open card sorting is ideal for exploring how users naturally categorize information without any predefined structure. It helps you create new information architectures, find patterns in user expectations, and generate ideas for structuring and labeling your app.

Closed Card Sorting

In closed card sorting, participants sort cards into predefined categories provided by the researcher. This makes it an evaluative research method, meaning it’s useful when you want to test the effectiveness of an existing information structure, but it won’t tell you how users naturally categorize concepts.

You can use closed card sorts to learn whether users understand your existing categories, identify misleading categories, and validate whether your information is presented the right way for your audience.

Hybrid Card Sorting

Hybrid card sorting combines elements of both open and closed card sorting. Participants can place cards into predefined categories or create new ones if they feel the existing options don’t fit. This method provides flexibility while still allowing you to test specific category structures.

Hybrid card sorts are great for validating existing information without closing the door on new ideas. This hybrid approach makes it a great tool for improving a live website or app.

Moderated vs. Unmoderated Card Sorting

Moderated card sorting involves a facilitator who guides the participants through the exercise. The facilitator is present to provide instructions, answer questions, and probe deeper into participants’ thought processes.

A moderator allows for real-time interaction and clarification. The facilitator can ask participants to explain their reasoning, which leads to richer insights. It’s useful when you need a deep understanding of the “why” behind users’ decisions. The downside is that it’s time-consuming, so lean teams or those with a “move fast” mindset don’t often opt for a moderated approach.

In unmoderated card sorting, participants complete the exercise on their own, typically using an online tool. Participants receive instructions and then sort the cards at their own pace without supervision.

An unmoderated card sort is more scalable and cost-effective. It’s convenient for participants, as they can complete the task at their own pace and on their own schedule. It also reduces the potential for facilitator bias.

The lack of a facilitator means there’s no opportunity to ask follow-up questions or clarify participants’ reasoning. As a result, you might miss out on insights into why participants sorted the cards in a particular way. Furthermore, without guidance, participants might misunderstand the task or make errors that a moderator could have corrected.

In-Person vs. Remote Card Sorting

You can conduct card sorting in person or remotely. Both methods are effective, but the choice depends on project needs and participant availability.

In-person card sorting, participants sort physical cards. This offers direct interaction, immediate feedback, and a tactile experience for the user. In some cases, people are better at categorization when they can physically move objects (note cards) around a table.

Remote card sorting, however, is the more common card sorting technique. It provides flexibility, scalability, and access to a broader audience. Researchers can conduct more sessions with a digital tool when they don’t need to meet in person. Remote card sorts are also less expensive.

When to Use Card Sorting

Card sorting is useful throughout various stages of the website and app design process. It helps you understand how users naturally group information. By revealing users’ mental models, it informs decisions that enhance the user experience, making it easier for visitors to find what they’re looking for.

This technique is particularly valuable when designing or redesigning navigation menus, categorizing products, or structuring complex information.

Additionally, card sorting plays a role in conversion funnel analysis. It can help identify where users might get confused or drop off due to poorly structured information, which is crucial for optimizing the funnel and improving conversions.

For example, if you observe a significant drop-off at a navigation step, card sorting can help redesign that menu to reduce confusion and keep users progressing toward conversion​.

How to Do Card Sorting (Step-by-Step)

By following these steps, you can effectively use card sorting to organize content to enhance user experience and support your broader goals, such as improving navigation, reducing bounce rates, and ultimately driving conversions.

Step 1: Define Your Goals

Your first step is to define what you hope to achieve with the card-sorting exercise. Are you trying to improve the navigation of a website? Categorize products? Understand user behavior? Understanding your goal will guide the rest of the process.

Step 2: Choose Open or Closed Card Sorting

Based on your goals, choose whether to conduct open or closed card sorting. Open sorting is great for exploration. Closed sorting is used to test predefined categories, and hybrid sorting combines both approaches.

Step 3: Select Your Participants

It’s important to choose participants that represent your target users. They should have a background, needs, and behaviors similar to those of your typical users.

This group may include a mix of people who have used your product in the past and potential users who are seeing it for the first time. The goal is to ensure that the results are relevant to the people who will use the site or app.

For qualitative insights, around 15-20 participants can provide meaningful patterns. For more robust statistical data, consider larger sample sizes.

Step 4: Prepare Your Materials

Write down each item, feature, or piece of content on individual cards. This can be done physically with actual cards or using online card sorting tools like Optimal Workshop or Miro. Label the cards clearly so all users can understand them.

How many cards should you create? It depends on your needs, but no more than 30-50 is about the sweet spot to prevent user fatigue.

If you’re doing a closed card sort, you’ll need to prepare the categories in advance. These categories should be mutually exclusive and cover all possible groups for the content. Label the cards as clearly as possible. You can use words, links, images, or even descriptions of pages or concepts.

Avoid using identical words for different cards, as users tend to group them automatically. For example, cards named Honda Civic, Honda CRV, and Honda Pilot are too similar. Users will likely put them into a Honda category, even if there are better ways to categorize them.

Step 5: Set Up the Sorting Exercise

Before participants sort cards, you’ll need to issue some instructions.

• Group the cards into whatever categories make the most sense to you.
• The number and size of groups can vary.
• You can rearrange or change your mind at any time.
• Leave it to the side if you don’t know what a card means.
• There are no right or wrong answers.
• Please think aloud and explain your thought process as you go.

If the exercise is in-person, you’ll need to explain the task carefully, whether they are creating their own groups (open sort) or sorting into predefined categories (closed sort).

If the exercise is remote, you’ll need to write out instructions as clearly as possible. Try to anticipate any question they might ask so you can answer it preemptively.

Step 6: Conduct the Card Sorting Exercise

Give each participant a set of unsorted cards. Randomize their order so there isn’t any inherent grouping. Ask participants to sort cards into whatever groups they find appropriate.

For instance, an ecommerce brand might give participants cards that say “T-shirts,” “hats,” “shoes,” “socks,” and “sweatpants.” A participant might put the “hats” card in a category called “Accessories.” Or they might put it in its own category.

Give the participants as long as they need to sort the cards. If you’re using an open card sorting exercise, ask them to label each group with a unique name, but only once they’ve finished sorting. It’s important that they do this last. Naming a group early can limit themselves to a specific category name.

If your exercise is in-person, observe the participants as they sort the cards, but avoid interfering. Note any comments or questions they have, as these can provide valuable insight into their thought process.

Most importantly, participants should be asked to verbalize their thought process as they sort the cards. This can reveal their reasoning and provide deeper insights into how they understand and organize the information.

Step 7: Ask Follow-up Questions

If you’re conducting the card sorting in person, take the opportunity to ask some follow-up questions. This is a great way to get inside their brains. Here are some questions you might ask, but feel free to come up with your own:

• Did any items seem like they should appear in multiple groups?
• Why did you choose those category labels? Did you consider other labels?
• Were any items especially hard to place?
• Why did you leave these items unsorted?

Step 8: Analyze the Results

After collecting the sorted data, look for patterns in how participants grouped the cards. In open sorting, focus on recurring themes and common groupings. In closed sorting, assess how well the predefined categories worked.

If you’re using card sorting software, comb through its analytics for insights. A particularly useful tool is a cluster analysis, which helps identify groups of cards that participants frequently place together.

Pay attention to any cards that were consistently placed in different categories by different users. These outliers may indicate unclear content or concepts that need further refinement.

Step 9: Implement and Test the Findings

Based on the results, design or refine your website or app’s information architecture. Ensure that the structure reflects the natural groupings identified during the card sort.

After implementing changes, it’s a good idea to conduct user testing to validate that the new structure works well in practice. This step helps you understand whether the restructured content improves the overall experience.

Step 10: Schedule Future Testing

Card sorting is not a one-time activity. Consider conducting additional card sorts as your content evolves to refine and adapt your information structure. The sites and apps that are best aligned with user needs go through cycles of continuous feedback and iteration.

Card Sorting Examples

Let’s look at three card sorting examples to help you understand their impact on the digital experience.

Mattel

The problem: Mattel wanted to review the information architecture of the Doll Showcase section of their Barbie collectors’ website. The Showcase was struggling with declining visits, likely due to the marketing-based organization of the site and layers of sub-navigation.

The goal: Learn more about the mental models that collectors use to explore the site and then reorganize the section’s architecture to match.

The card sort: Mattel ran two card sorting methods: 1) An open sort of cards from a fashion section of the Showcase, and 2) A closed sort of cards that represented the existing architecture.

The results: They learned that collectors don’t have a universal mental model. They browse in their own ways, so a single search method isn’t sufficient. They ultimately created a faceted navigation system that allowed searching by doll name, serial numbers, themes, and release year. As a result, the use of Showcase increased dramatically.

The Telegraph

The Telegraph is one of our clients. We worked with them to help optimize their subscriber experience.

The problem: The Telegraph was struggling to convert readers into paying subscribers. Their varied audience means one-size-fits-all solutions aren’t effective. Plus, A/B testing on a live site of their size is expensive.

The goal: Improve the paywall experience (the “subscribe to read” call-to-action). They also wanted to improve the user experience for readers to help them find the content they need.

The card sort: We used card sorting in conjunction with other rapid testing tools. This allowed us to identify opportunities for live tests that had a high level of success.

The results: We helped The Telegraph increase paywall conversions, keep those new subscribers, and improve content exploration throughout the site. Most importantly, we shifted their internal mindset by showing them how to reach customers to get feedback without disrupting their usual operations.

Singapore Polytechnic

The problem: Singapore Polytechnic’s website content was scattered and chaotic. The site included confusing layers of sub-navigation, and the university’s schools all used different branding. As a result, content was hard to find.

The goal: Understand how users group information and learn where they expect to find content, then build an information architecture that provides it.

The card sort: The optimization team used a digital card sorting tool to conduct an open card sorting study. Participants were asked to sort cards of the existing content into whatever categories they saw fit. The data was later narrowed down using tree testing.

The results: Researchers identified three major content groups: student life, courses (grouped by topic), and admissions and financial matters. This formed the foundation of the architecture for Singapore Polytechnic’s new site.

Frequently Asked Questions About Card Sorting

Here are some common questions people ask about card sorting that might help you understand the concept better.

Is Card Sorting Qualitative or Quantitative?

Card sorting can be both qualitative and quantitative. Qualitative insights come from understanding why participants group items a certain way, while quantitative data is gathered by analyzing patterns and frequencies in groupings across many participants.

How Long Does a Card Sorting Session Take?

A card sorting session typically takes 15 to 45 minutes, depending on the number of cards and the complexity of the task. For the best responses from your participants, keeping the exercise as short as possible is best.

Are There Any Disadvantages to Card Sorting?

Disadvantages include potential participant bias, especially in moderated sessions, and the time-consuming nature of analyzing qualitative data. Additionally, the results may not always translate directly into an effective information architecture.

What is Card Sorting in UX?

Card sorting in UX is a research technique where users organize content into groups, revealing their mental models. This helps designers create intuitive navigation structures and improve the overall user experience by aligning the website or app layout with user expectations.

What is Card Sorting in Project Management?

In project management, card sorting is used to prioritize tasks, organize project elements, or brainstorm ideas. It’s a visual way for teams to identify patterns, set priorities, and ensure everyone has a shared understanding of the project’s structure.

What is Card Sorting in Website Design?

Card sorting in website design involves users grouping content or features to reveal their preferred organization. This process guides the design of navigation menus, category names, and information architecture. The goal is to make a website that meets user expectations and is, therefore, easy to navigate.

Card Sorting and Your Broader Toolkit

Card sorting is a valuable tool in your broader UX and optimization toolkit, but it’s most effective when used in conjunction with other methods. It’s one tool of our 5 Factors framework, a set of competencies that set high-performance teams apart.

By integrating card sorting into this broader toolkit, you can ensure that your site not only meets user expectations but also performs well in areas that matter most, such as navigation, content findability, and overall user experience.

But card sorting is just the beginning. To truly optimize your digital product, it’s critical that you build an optimization program that includes all of the right qualities. Evaluate how your optimization efforts stack up against the top performers by taking this short 5-Factors assessment. It will show you where to invest your resources to improve your overall digital experience.